Naoto Iwase
Search, experience reuse, and evaluation in agents that conduct machine learning experiments
Comparing ways to improve the systems around models through search, learning, and evaluation
A survey of small networks that reason by test-time recursion
Systematizing the signals and methods that make LLM reasoning reliable
Systematic survey of Diffusion Language Model literature (formulation through applications)
The frontier of single-step generative models (from Flow Matching to Drifting Models)
A fully open Vision-Language Model with video grounding capabilities
Fully open language and reasoning models at 7B/32B scale (full Model Flow released)